GEOPHYSICAL RESEARCH LETTERS, VOL. 35, L12505, doi:10.1029/2008GL033365, 2008 Click Here for Full Article Changes in West Antarctic ice stream dynamics observed with ALOS PALSAR data Eric Rignot1,2 Received 21 January 2008; revised 27 February 2008; accepted 10 March 2008; published 28 June 2008. [1] The Advanced Land Observation System (ALOS) challenging to get good coherence in certain parts of Phased-Array Synthetic-Aperture Radar (PALSAR) is an Antarctica, except for a brief period of time in 1996 when L-band frequency (1.27 GHz) radar capable of continental- two Earth Remote Sensing (ERS) satellites were put along scale interferometric observations of ice sheet motion. Here, the same orbit one day apart, and in winter 1992 and 1994 we show that PALSAR data yield excellent measurements when the ERS-1 satellite was in a 3-day exact repeat mode. of ice motion compared to C-band (5.6 GHz) radar data The opportunity for short-term mapping of wet, dynamic because of greater temporal coherence over snow and firn. coastal regions of Antarctica disappeared in March 2000 We compare PALSAR velocities from year 2006 in Pine when ERS-1 ceased operations. Unfortunately, these sectors Island Bay, West Antarctica with those spanning years 1974 are the most important for the present-day ice sheet mass to 2007. Between 1996 and 2007, Pine Island Glacier sped balance [Rignot et al., 2008]. up 42% and ungrounded over most of its ice plain. Smith [4] Here, we examine data collected by the ALOS Glacier accelerated 83% and ungrounded as well. Their PALSAR L-band radar in 2006, six months after launch. largest speed up are recorded in 2007. Thwaites Glacier is We show results from the Amundsen Sea sector of West not accelerating but widening with time and its eastern ice Antarctica and compare them with those obtained with shelf doubled its speed. Total ice discharge from these Radarsat, ERS, and Landsat between 1974 and 2007. glaciers increased 30% in 12 yr and the net mass loss Changes in flow speed measured at the grounding lines increased 170% from 39 ± 15 Gt/yr to 105 ± 27 Gt/yr. are converted into changes in mass flux, which in turn are Longer-term velocity changes suggest only a moderate converted into changes in mass balance. We conclude with a loss in the 1970s. As the glaciers unground into the discussion of the evolution of glaciers and mass balance in deeper, smoother beds inland, the mass loss from this this region and their past and future impact on sea level rise. region will grow considerably larger in years to come. Citation: Rignot, E. (2008), Changes in West Antarctic ice 2. Data and Methodology stream dynamics observed with ALOS PALSAR data, Geophys. Res. Lett., 35, L12505, doi:10.1029/2008GL033365. [5] The Advanced Land Observation System (ALOS) Phased-array L-band Synthetic-Aperture Radar (PALSAR) was developed by the Japan Aerospace Exploration Agency 1. Introduction (JAXA) and the Japanese Resources Observation System [2] Synthetic-aperture radar interferometry (InSAR) has Organization (JAROS) and launched in January 2006. The proven to be a valuable tool for measuring ice motion and radar operates at 1.27 GHz (24 cm wavelength). The data grounding line position [e.g., Joughin et al., 1996; Rignot, analyzed herein were acquired in May–September 2006, at 1996]. Estimates of ice sheet mass balance have been horizontal receive and transmit polarization, with a 41.5° obtained with this technique for the entire Greenland and look angle off nadir, 7.5 m  4.1 m resolution, respectively, Antarctic ice sheets, with few areas left unsurveyed [Rignot in slant-range (across-track) and azimuth (along-track), a and Kanagaratnam, 2006; Rignotetal., 2008]. Most 70 km swath width, a À34 dB noise equivalent backscatter InSAR data collected to date have been acquired at the cross section, a 46-day repeat cycle, and sun-synchronous, C-band frequency (5.6 GHz or 5.6 cm wavelength). Data ascending (evening) orbits of the satellite. The data were collected at the L- and C-band frequencies by the Space obtained as raw individual frames from JAXA via the Alaska Shuttle Experiment SIR-C in 1994, however, indicated that Satellite Facility (ASF). Frames were concatenated and longer radar wavelengths offer superior temporal coherence processed into single look complex tracks using the Gamma on snow and ice surfaces because of their penetration into the Remote Sensing processor (http://www.gamma-rs.ch). snow and firn [Rignot et al., 1996, 2001; Dall et al., 2001]. [6] A speckle tracking technique [Michel and Rignot, Yet, few satellite L-band InSAR data have been collected on 1999] was applied on image pairs acquired 46 days apart. glacier ice, and nearly none on ice sheets. Averaging boxes for speckle tracking were 128 (range)  [3] C-band radars have been successful at measuring ice 256 (azimuth) samples in size, with a grid spacing of 32  motion even though they were not designed for this appli- 64 samples, and a search window of 64  64 samples. The cation, but their month-long repeat cycle has made it offsets were median filtered to remove bad matches. A plane fit was adjusted through non-moving areas such as ice 1 islands, nunataks, and ice caps and removed from the offset Earth System Science, University of California, Irvine, California, fields to obtain absolute displacements. The displacements USA. 2Jet Propulsion Laboratory, Pasadena, California, USA. were then converted into easting and northing velocities assuming surface parallel flow and using a digital elevation Copyright 2008 by the American Geophysical Union. model of Antarctica that combines ICESAT and ERS-1 0094-8276/08/2008GL033365$05.00 L12505 1of5 L12505 RIGNOT: ALOS PALSAR ANTARCTICA L12505 Figure 1. Velocity of West Antarctic glaciers draining into the Amundsen Sea using (a) ERS-1/2 January–March 1996, (b) PALSAR May–September 2006; and (c) Radarsat-1 August–September 2006 data overlaid on radar brightness color coded on a logarithmic scale from 10 m/yr (brown), to 50 m/yr (green), 100 m/yr (blue) to 1000 m/yr (pink) and more than 1000 m/yr (red). Grounding line positions for year 1992 and 1996 are black thin lines along the coast. radar altimetry (J. Bamber, unpublished data, 2006) climate model [van den Broeke et al., 2006; van de Berg et (Figure 1a). The precision of PALSAR velocities is 2 m/yr, al., 2006] and averaged for the years 1980–2004. We or 3 times better than Radarsat-1. estimated the absolute precision of the accumulation rates [7] We compare the PALSAR velocities with those to be 14% in this sector [Rignot et al., 2008]. measured using ERS-1/2 ascending/descending tracks in spring 1996 (5 m/yr precision); Radarsat-1 in years 2000 3. Results to 2005 (6 m/yr precision); and Landsat imagery [Lucchitta et al., 1994; Ferrigno et al., 1993] (5 to 10% precision). To [10] PALSAR 46-day speckle tracking (Figure 1b) works do the latter, we visually placed the Landsat measurement well even in areas where signal coherence is lost with points in the PALSAR image mosaic as no precise latitude Radarsat-1 (Figure 1c), which makes velocity calibration longitude information was available for these data. The easier and more robust with PALSAR. Signal coherence is error associated with this process is small compared to the the multiplicative product of the decorrelation from the inherent precision of the Landsat measurements. interferometric baseline, volume scattering from snow, [8] Ice fluxes are calculated by combining surface velocity thermal noise from the radar system and changes in the with ice thickness data from year 2002 [Thomas et al., reflecting surface [Hoen and Zebker, 2000]. Radarsat-1 2004a]. Ice velocity is mapped with no data gap along the results on other years/seasons, shorter/longer baselines, are selected thickness profiles in the study area only for years similar to those shown here, so the loss of coherence is not 1996 and 2006. To calculate ice fluxes in other years, we due to the baseline or snow scattering. The radar backscatter determined the multiplicative factor to apply on the 2006 of the snowy surface is high above the noise floor, so velocities to match the other year velocities and applied the decorrelation from thermal noise is negligible as well. This same factor on the 2006 mass flux. This is justified by the leaves surface weathering as the main cause for the loss of fact that the multiple year velocity profiles are brought into coherence after 24 days (Figure 1c) versus 1-day (Figure 1a). excellent agreement with the scaling technique. We neglect [11] At L-band, signal coherence is high on ice shelves, glacier thinning before or after the year of the thickness which are radar-bright because of percolation and refreezing measurements (2002) because glacier thinning of 2 m/yr at of melted snow. The longer repeat cycle and higher look the location of the thickness data [Thomas et al., 2004a] angle of PALSAR versus Radarsat-1/ERS-1 causes vertical introduces an uncertainty of +12 to À10 m for years 1996 displacements of the ice shelf induced by oceanic tides (1 m) and 2007, respectively, which is 1% of a total thickness of to be much smaller than horizontal ice motion in 46 days 1.2 km. We have no information on thinning rates prior to (140 to 380 m). Hence, PALSAR-derived ice-shelf veloci- 1992, but they were presumably smaller than for 1974– ties have little tidal contamination. At high elevation and on 2002 because the glacier speed up was less.